PrivacyHigh
Meta Allegedly Used Books3, a Dataset of 191,000 Pirated Books, to Train LLaMA AI
- Occurred
- Oct 25, 2020
- Published
- Jul 19, 2025
- Source
- AI Incident Database
Summary
Meta and Bloomberg allegedly used Books3, a dataset containing 191,000 pirated books, to train their AI models, including LLaMA and BloombergGPT, without author consent. Lawsuits from authors such as Sarah Silverman and Michael Chabon claim this constitutes copyright infringement. Books3 includes works from major publishers like Penguin Random House and HarperCollins. Meta argues its AI outputs are not "substantially similar" to the original books, but legal challenges continue.
Developers
- The Pile
- Shawn Presser
- Meta
- EleutherAI
- Bloomberg
- Generative AI developers
Deployers
- Meta
- EleutherAI
- Bloomberg
- Generative AI developers
Harmed parties
- Zadie Smith
- Writers
- Verso
- Stephen King
- Sarah Silverman
- Richard Kadrey
- Publishers found in Books3
- Penguin Random House
- Oxford University Press
- Over 170,000 authors found in Books3
- Michael Pollan
- Margaret Atwood
- Macmillan
- HarperCollins
- General public
- Creative industries
- Christopher Golden
- Authors
Tags
- Copyright
- Copyright infringement
- Large language model
- Training data
- Pirated books
- Piracy
- Intellectual property theft
- Lawsuits
- Fair use
- Unauthorized AI training
- Books
- Model transparency
- Data scraping
- Publishing
- Legal action
Classification
Category confidence: 85%
Severity confidence: 90%
Rationale: Unauthorized use of copyrighted material for AI training constitutes privacy and consent violation with active lawsuits.
Classifier: Qwen/Qwen3.5-35B-A3B-FP8